Macam Macam Pelarut

Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 12

Nhien et al.

Biotechnol Biofuels (2017) 10:81


DOI 10.1186/s13068-017-0767-3 Biotechnology for Biofuels

RESEARCH Open Access

Techno-economic assessment of hybrid


extraction and distillation processes for furfural
production from lignocellulosic biomass
Le Cao Nhien1†, Nguyen Van Duc Long1†, Sangyong Kim2 and Moonyong Lee1*

Abstract
Background: Lignocellulosic biomass is one of the most promising alternatives for replacing mineral resources to
overcome global warming, which has become the most important environmental issue in recent years. Furfural was
listed by the National Renewable Energy Laboratory as one of the top 30 potential chemicals arising from biomass.
However, the current production of furfural is energy intensive and uses inefficient technology. Thus, a hybrid purifica-
tion process that combines extraction and distillation to produce furfural from lignocellulosic biomass was considered
and investigated in detail to improve the process efficiency. This effective hybrid process depends on the extracting
solvent, which was selected based on a comprehensive procedure that ranged from solvent screening to complete
process design.
Results: Various solvents were first evaluated in terms of their extraction ability. Then, the most promising solvents
were selected to study the separation feasibility. Eventually, processes that used the three best solvents (toluene,
benzene, and butyl chloride) were designed and optimized in detail using Aspen Plus. Sustainability analysis was
performed to evaluate these processes in terms of their energy requirements, total annual costs (TAC), and carbon
dioxide (CO2) emissions. The results showed that butyl chloride was the most suitable solvent for the hybrid furfural
process because it could save 44.7% of the TAC while reducing the CO2 emissions by 45.5% compared to the toluene
process. In comparison with the traditional purification process using distillation, this suggested hybrid extraction/
distillation process can save up to 19.2% of the TAC and reduce 58.3% total annual CO2 emissions. Furthermore, a
sensitivity analysis of the feed composition and its effect on the performance of the proposed hybrid system was
conducted.
Conclusions: Butyl chloride was found to be the most suitable solvent for the hybrid extraction/distillation process
of furfural production. The proposed hybrid sequence was more favorable than the traditional distillation process
when the methanol fraction of the feed stream was <3% and more benefit could be obtained when that fraction
decreased.
Keywords: Biomass feedstock, Biorefinery development, Furfural, Hybrid process, Lignocellulosic biomass,
Solvent selection

*Correspondence: mynlee@yu.ac.kr

Le Cao Nhien and Nguyen Van Duc Long contributed
equally to this work
1
School of Chemical Engineering, Yeungnam University,
Gyeongsan 712-749, South Korea
Full list of author information is available at the end of the article

© The Author(s) 2017. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License
(http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium,
provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license,
and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/
publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Nhien et al. Biotechnol Biofuels (2017) 10:81 Page 2 of 12

Background production process includes two main sections: reaction


In recent years, the interest in renewable resources has and puriication. After pretreatment, a pentosane-rich
increased considerably owing to environmental problems biomass consisting of a material such as corn cobs or sug-
and the overdependence on mineral resources. Lignocel- arcane bagasse is introduced to a series of reactors to be
lulosic biomass, which is the most abundant feedstock hydrolyzed to pentose, which is then dehydrated to fur-
on the Earth, is a readily renewable resource for replac- fural, using sulfuric acid as the catalyst. Subsequently, the
ing fossil fuels. he use of this biomass feedstock for the vapor stream from the reactor, which consists of furfural
manufacture of biochemicals and biofuels can gradually (about 6 wt%), several byproducts (4 wt%), and water
replace fossil-based feedstock. he conversion of biomass (making up the balance), is normally liqueied to make
into chemicals conceptually brings a promise of sustain- secondary steam before being puriied using a distilla-
able, inherently safer, and eco-friendly production [1]. tion technique. In the reaction section, the low reactivity
Furfural, which is one of the top 30 potential chemicals and poor mass transfer lead to a molar yield of furfural of
arising from biomass, is a key bio-based platform chemi- about 50% in today’s process [3].
cal that can be used to replace oil-based chemicals [1]. Meanwhile, in the puriication part, removing a large
Currently, its global production capacity is approximately quantity of water by distillation is very energy intensive
300 ktons/year, and it is primarily used for producing because of the presence of a heterogeneous azeotrope
herbicides, stabilizers, pharmaceuticals, and numerous between furfural and water (35.46 wt% furfural) [4, 5].
resins [2]. Zeitsh presented a comprehensive overview Note that this process is still used to produce approxi-
of traditional furfural technology and recent studies on mately 80% of the furfural supply on today’s world
this process [3]. Although the irst industrial production market because of its low capital investment, easy imple-
of furfural was by the Quaker Oats Company in the early mentation, and inexpensive raw materials [6]. he reac-
1920s in Iowa, the current technologies used for furfural tion and separation sections account for major portions
production have not been improved signiicantly [3]. Fig- of the total production costs. herefore, improvements in
ure 1 shows a block low diagram of the furfural produc- these steps will lead to more cost-efective and sustain-
tion from lignocellulosic biomass. he typical furfural able processes.
In recent years, several processes have been reported
in the literature and pilot-scale furfural production has
been conducted, including the Supray Yield process,
Lignocellulosic Biomass work of Vedernikovs, CIMV in France, Bioreinery by
the Lignol Innovations Corporation in British Columbia,
(corncobs, oat hulls, and multi-turbine column processes [3, 7–9]. Remark-
bagasse…) ably, these processes focused only on improving the
furfural yield in the reaction part, without giving much
attention to the energy-intensive puriication part. A
heat pump technique was applied in the pentose-to-fur-
Pretreatment fural process to recover heat from the tops of the reactor
and distillation column [10]. However, the use of a com-
pressor might encounter several serious issues that could
limit the use of a heat pump in the industry, such as the
occurrence of overheating at a high operating tempera-
Acidic Hydrolysis ture, high capital cost, and process complexity when
operating the compressor equipment. Another inves-
tigation on the furfural puriication process involved
the use of a combined integrated and intensiied tech-
nique, which showed a decrease in the total annual
Purification cost of about 10% [4]. Although distillation techniques
require relatively high amounts of energy, they are still
the most commonly used separation methods for purify-
ing furfural on an industrial scale because of their ease
Furfural >99 wt% of implementation and large capacity. It is worth noting
that in the distillation process the presence of organic
Fig. 1 Schematic diagram of the furfural production from lignocel- acids in the luid at a high temperature will result in the
lulosic biomass loss of furfural [7].
Nhien et al. Biotechnol Biofuels (2017) 10:81 Page 3 of 12

On the other hand, liquid–liquid (LL) extraction may Methods


be more promising from the energy viewpoint. How- Systematic procedure for solvent selection
ever, inding an efective solvent and designing a solvent he selection of an efective solvent is critical to the
regeneration part make the use of an extraction method design of the hybrid extraction/distillation process. he
more challenging. he solvent selected severely afects solvent should not only have a high equilibrium distri-
the performance of the extraction process. he solvent bution and speciic selectivity but also provide easy sol-
must be satisfactory in terms of its cost, selectivity, dis- vent regeneration and economic feasibility. In this study,
tribution coeicient, density, viscosity, toxicity, boiling the solvents were selected using a systematic procedure
point temperature, easy regeneration, etc. that was well described in our previous study [16]. As
Several methods for selecting extraction solvents shown in Fig. 2, the comprehensive procedure for select-
have been well described in the literature, such as the ing a solvent for the production of furfural has essentially
tradition experimental method, computer-aided molec- six steps. First, several preliminary simulations are car-
ular design (CAMD) [11–15], and the combination of ried out to select promising solvents from the literature.
solvent screening and process design [16]. Carrying he selected solvents are then screened on the basis of
out multiple experiments to ind a suitable solvent pro- their selectivity and equilibrium distribution. Next, azeo-
vides relatively accurate and reliable results, but there trope investigations and the ease of solvent regeneration
are clear limitations in terms of the time, cost, and a are explored in the separation feasibility step. Processes
number of solvents tested. CAMD methods, on the using the most promising solvents are eventually
other hand, can screen a large number of structural designed and optimized in detail. he total annual costs
molecules and determine the molecule, which matches (TAC) and carbon dioxide (CO2) emissions of all the pro-
the target properties. In particular, a general CAMD cesses are calculated to make a fair comparison, and the
can be formulated as a mixed-integer nonlinear pro-
gramming (MINLP), in which several studies proposed
the solution successfully [17, 18]. Tula et al. employed a
CAMD approach to process design [11]. Note that if a
solvent is selected without designing the entire process, Generating solvent alternatives
including extraction and solvent regeneration units, it Literature studies, preliminary
is not possible to evaluate the economic feasibility of simulations
the process. In contrast, a method that combines sol-
vent screening and the design of the entire process can
Solvent screening
assess both the extraction abilities of the solvents and
Selectivity, equilibrium distributions
the economic feasibility and sustainability of the sol-
vent processes [16].
In this study, hybrid extraction and distillation pro- Separation feasibility
cesses for furfural production from lignocellulosic Azeotrope investigation, solvent
biomass were designed and optimized through a compre- regeneration
hensive framework for solvent selection. Numerous sol-
vents were initially screened on the basis of a literature
review and preliminary simulations. Next, the separa- Process design
tion feasibility of each potential solvent was considered Detailed design, optimization
before designing and optimizing the processes for the
most promising solvents. All of the processes were simu-
lated using Aspen Plus V9 and were assessed in terms of Evaluation
their energy requirement, economic performance, and Energy, costs, carbon footprint
environmental impact to make a fair comparison. Fur-
thermore, the proposed hybrid process with the selected
solvent was compared with the distillation puriication Proposed solvents
process. In the lignocellulose-based process, feed compo- Overall performance
sition uncertainties are inherent and may have an adverse
Fig. 2 Systematic procedure for selecting solvent for furfural produc-
efect on the suggested optimum. herefore, a sensitivity
tion
analysis was also conducted.
Nhien et al. Biotechnol Biofuels (2017) 10:81 Page 4 of 12

most promising solvents are accordingly proposed. he solvents used have purities of more than 99 wt%.
potential solvents and detailed designs for producing T he solvent and feed streams are initially introduced
furfural on an industrial scale can be investigated using to an LL extractor to produce a solvent-rich stream
a framework. called the extract and an extracted-feed stream called
the rainate. The extract that contains the most furfural
Simulation in the feed is input to a distillation column to purify
Aspen Plus V9 was used to simulate the puriication pro- the furfural and recover the solvent. he following
cess for furfural production utilizing the RadFrac model sections show how diferent solvents were evaluated
for rigorous distillation and Extraction model for the LL for the hybrid extrac- tion/distillation process using the
extraction. he non-random two-liquid (NRTL) prop- systematic framework shown in Fig. 2.
erty method with binary parameters taken from the liq-
uid–liquid equilibrium (LLE) Aspen database was used to Solvent screening
calculate the liquid activity coeicients for the Extraction Because the art of solvent extraction has a long history,
model. Meanwhile, the non-random two-liquid-Hayden– and LL extraction has been practiced in a large number
O’Connell (NRTL-HOC) thermodynamic package with of industrial applications, the key principles of solvent
binary parameters taken from the vapor–liquid equilib- selection are available and well described in the literature
rium (VLE) Aspen database was used to perform a rigor- [20–22]. Moreover, several researchers investigated the
ous simulation of the distillation, in which the NRTL was extraction of furfural from an aqueous stream using the
used to calculate the liquid activity coeicients, and the acidic hydrolysis of biomass [4, 23]. On the basis of lit-
HOC equation of state was used to calculate the thermo- erature survey and heuristics, all possible solvents from a
dynamic properties of the vapor phase. he HOC equa- wide range of chemical families, including alcohols, nor-
tion reliably predicts a mixture of carboxylic acids, which mal hydrocarbons, ketones, aromatics, amines, etc., are
may occur during the solvation of polar compounds irstly generated as shown in Fig. 3. Subsequently, several
and dimerization in the vapor phase [19]. Note that the preliminary simulations are carried out in order to pre-
interaction parameters between the major components, pare a short list of candidates for detail screening step.
such as furfural–water and all solvents–water, are avail- In particular, ten potential solvents (toluene, benzene,
able in both the LLE and VLE Aspen databases. For the p-xylene, octyl acetate, decane, cyclohexane, hexene,
missing binary parameters of the remaining minor com- cyclohexene, cumene, and butyl chloride) were selected
ponents, a Universal Quasichemical Functional-Group for further consideration.
Activity Coeicients (UNIFAC) model was employed for he LL equilibrium curve plays a crucial role in design-
estimation. ing an extraction process [24]. he use of equilibrium
distribution coeicients that can screen a large number
Results and discussion of solvents is an efective and economic method for sol-
In this study, the furfural puriication process was vent screening [25]. he distribution coeicient was cal-
designed based on a production rate of 50 ktons/year. culated using the equilibrium data at a given temperature
he feed stream consisted of aqueous solutions obtained as follows [23]:
from the acidic hydrolysis reaction of the lignocellulosic
KD = [Extracted component]solvent [Extracted component]aqueous.
biomass. Table 1 lists the feed conditions, component
compositions, and product speciications [3]. All of the (1)

Table 1 Feed mixture conditions and product speciica- he distribution coeicient of the extracted compo-
tions nent (KD) is the ratio of the weight percentage (wt%) of
Component Mass fraction (wt%) the extracted component [furfural, methanol, or acetic
acid (AA)] in the solvent phase to the weight percentage
Methanol 2.0 of the extracted component in the aqueous phase. In the
Water 90.0 solvent screening step, the water–solvent equilibrium for
Acetic acid (AA) 2.0 furfural, AA, and methanol was studied, and the distri-
Furfural 6.0 bution coeicients and extraction eiciencies were com-
Furfural product purity (wt%) >99.0 pared for each solvent at 298 K. he feed low rate was
Solvent product purity (wt%) >99.0 ixed, and the solvent low rate was changed to analyze
Temperature (K) 353 the solvent eiciency. In particular, for each solvent, four
Pressure (kPa) 101 diferent cases with feed-to-solvent mass ratios of 1:0.4,
Mass flowrate (kg/h) 105,000 1:0.9, 1:1, and 1:1.1 were considered.
Nhien et al. Biotechnol Biofuels (2017) 10:81 Page 5 of 12

Fig. 3 Lists of solvents: from possible initial solvents to short list of candidates for detailed screening

Figure 4 shows the equilibrium curves of the furfural an excellent equilibrium with a very high AA concentra-
for the diferent solvents examined at 298 K. Clearly, ben- tion in the extract phase, whereas all of the remaining
zene, toluene, p-xylene, cumene, octyl acetate, and butyl solvents have unfavorable equilibrium for AA extraction.
chloride have a favorable equilibrium with high distri- In particular, with the exception of decane, the presence
bution coeicient values, whereas decane, cyclohexane, of AA in the rainate phase was always much greater
cyclohexene, and hexane show unfavorable equilibrium than that in the extract phase.
for extracting furfural. Remarkably, benzene presents Figure 6 shows the equilibrium curves of methanol for
the most favorable equilibrium with the highest distribu- the diferent solvents examined at 298 K. Although octyl
tion coeicient value for furfural extraction. Accordingly, acetate shows the most favorable result for methanol
decane, cyclohexane, hexane, and cyclohexene were extraction, the methanol concentration in the rainate
deemed unsuitable for extracting furfural from an aque- phase was much higher than that in the extract phase.
ous solution. As a result, all of the solvents tested were unsuitable for
Figure 5 presents the equilibrium curves of AA for all extracting methanol from the aqueous feed stream.
the solvents tested at 298 K. Interestingly, decane shows

Fig. 5 Equilibrium curves of acetic acid using different solvents at


Fig. 4 Equilibrium curves of furfural using different solvents at 298 K 298 K
Nhien et al. Biotechnol Biofuels (2017) 10:81 Page 6 of 12

for furfural extraction compared to the other solvents.


For AA extraction, only decane was promising, with an
extremely high coeicient of AA. All of the solvents were
deemed unsuitable for extracting methanol from the
aqueous solution. It is worth noting that in the furfural
production process, AA and methanol were treated as

Table 2 Equilibrium distribution coeicient of furfural,


AA, and methanol for diferent solvents at 298 K
Solvent Feed-to-solvent KD
ratio (mass)
Furfural AA Methanol

Toluene 1:0.4 37.33 0.12 0.04


Fig. 6 Equilibrium curves of methanol using different solvents at 1:0.9 41.11 0.12 0.04
298 K 1:1 44.89 0.12 0.04
1:1.1 48.68 0.12 0.04
Benzene 1:0.4 66.57 0.16 0.04
In addition to the extraction ability, the amount of 1:0.9 73.76 0.16 0.04
water in the extract also has a great impact on the 1:1 80.99 0.17 0.04
choice of solvent. If the extract contains a large quantity 1:1.1 88.25 0.17 0.04
of water, greater efort for the subsequent distillation is p-Xylene 1:0.4 23.50 0.09 0.04
needed to achieve the desired furfural purity, consider- 1:0.9 25.73 0.09 0.04
ing the heterogeneous azeotrope between the furfural 1:1 27.97 0.09 0.04
and water. Figure 7 presents the amount of water in the 1:1.1 30.22 0.09 0.04
extract for diferent solvents at a feed-to-solvent ratio Octyl acetate 1:0.4 13.46 0.26 0.10
of 1:1. he octyl acetate shows the highest water low of 1:0.9 14.47 0.26 0.10
about 600 kg/h, while the extracts of other solvents con- 1:1 15.50 0.26 0.09
tain very small amounts of water. In the case of octyl ace- 1:1.1 16.53 0.26 0.09
tate, because 600 kg/h accounts for only 0.6 wt% of water Decane 1:0.4 4.05 0.10
1E+07
in the feed, it was still considered for the next step. 1:0.9 4.21 0.10
5E+07
Based on the equilibrium data, the equilibrium distri- 1:1 4.38 0.09
2E+08
bution coeicients of furfural, AA, and methanol for all 1:1.1 4.55 0.09
8E+08
the solvents at 298 K were calculated as listed in Table 2. Cyclohexane 1:0.4 2.04 0.07 0.02
Cyclohexane, hexane, and cyclohexane, which had very 1:0.9 2.15 0.07 0.02
low coeicients of furfural, were deemed unsuitable 1:1 2.27 0.07 0.02
1:1.1 2.39 0.07 0.02
Hexene 1:0.4 1.84 0.14 0.03
1:0.9 1.92 0.14 0.03
1:1 2.00 0.14 0.03
1:1.1 2.08 0.14 0.03
Cyclohexane 1:0.4 2.05 0.12 0.02
1:0.9 2.13 0.12 0.02
1:1 2.22 0.12 0.02
1:1.1 2.30 0.12 0.02
Cumene 1:0.4 16.85 0.10 0.03
1:0.9 18.31 0.10 0.03
1:1 19.77 0.10 0.03
1:1.1 21.24 0.10 0.03
Butyl chloride 1:0.4 15.20 0.13 0.04
1:0.9 16.34 0.13 0.04
Fig. 7 Water flow rate in extract for different solvents at feed-to- 1:1 17.50 0.13 0.04
solvent ratio of 1:1 1:1.1 18.66 0.13 0.04
Nhien et al. Biotechnol Biofuels (2017) 10:81 Page 7 of 12

valuable byproducts. Hence, a solvent’s extraction ability acetate was excluded from further consideration. Overall,
for them has less impact than for furfural. In the case of toluene, benzene, and butyl chloride were selected for the
decane, in addition to its excellent ability to extract AA, next process design step.
its performance in extracting furfural was still promising.
herefore, it was not eliminated in this step. Overall, ben- Process design
zene, toluene, p-xylene, cumene, octyl acetate, butyl chlo- In this step, the hybrid extraction/distillation processes
ride, and decane were selected for further consideration. for furfural production using the three best solvents
(toluene, benzene, and butyl chloride) selected from the
Separation feasibility previous step were designed and optimized. he aque-
If the added solvent forms azeotropes with the major ous feed stream, which was cooled down to 313 K, and
extracted components, the design of the solvent regen- the solvent were irst input to an extractor to produce an
eration part will become much more complex, lead- extract composed of a large portion of furfural in the feed
ing to economic infeasibility. herefore, in this step, the and a rainate. he extract was then introduced to a dis-
seven solvents selected in the solvent screening step were tillation column to separate the furfural and solvent. Note
investigated to determine whether they formed azeo- that in this hybrid process, the rainate from the extrac-
tropes with the feed components to examine the ease of tor consists of water loaded with AA and methanol which
separation. Originally, furfural and water form a binary are delivered into a wastewater treatment plant. In par-
heterogeneous azeotrope (35.5 wt% furfural) at 371 K ticular, AA is not collected in both hybrid and traditional
and 101 kPa. Table 3 lists all of the azeotropes formed distillation processes while methanol is a by-product in
between the added solvents and feed components, the distillation process. Comparison of both processes
which were investigated based on the Aspen database will be discussed further in sustainability analysis section.
at a pressure of 101 kPa. Because the major task of the For design speciications, all of the design variables such
hybrid extraction/distillation process is the separation as a total number of trays, feed location, and feed-to-sol-
of the furfural and solvent mixture, forming an azeo- vent ratio were manipulated through sensitivity analyses
trope between the furfural and solvent will lead to either to improve the process eiciency while maintaining the
infeasible separation or economic infeasibility. herefore, product purities and recoveries. To make a fair com-
the p-xylene, decane, and cumene solvents were elimi- parison, the furfural recovery was 99.9 wt% through the
nated because they formed homogeneous azeotropes extractor and 99.0 wt% through the distillation column,
with furfural. Note that in all cases except octyl acetate, resulting in the same furfural production rate in all cases.
the extracts contained a very small amount of water (less
than 100 kg/h). hus, the efect of the heterogeneous aze- Toluene solvent process
otropes formed between the water and solvents was neg- he feed stream obtained from the biomass hydroly-
ligible. However, the extract of the octyl acetate process sis process was irst cooled down to 313 K before being
contained quite a large amount of water (600 kg/h when introduced to the extractor (E1). Herein, the LL extrac-
the feed/solvent mass ratio was 1:1). he fact that water tion process produced a rainate at the bottom and an
not only forms a heterogeneous azeotrope with furfural extract containing furfural at the top. he extract was
but also with octyl acetate will make the following distil- then inputted to a distillation column (D1) to deliver a
lation process more complex and costly. herefore, octyl top stream of toluene and a bottom stream of the desired

Table 3 Investigation of azeotropes between feed components and diferent solvents


Azeotrope Solvent
Toluene Benzene p-Xylene Octyl acetate Decane Cumene Butyl chloride

Furfural-solvent – – Homogeneous – Homogeneous Homogeneous –


(97.1% p-xylene) (45.5% decane) (66.1% cumene)
Water–solvent Heterogeneous Heterogeneous Heterogeneous Heterogeneous Heterogeneous Heterogeneous Heterogeneous
(80.5% toluene) (91.1% benzene) (34.3% p-xylene) (16.4% octyl (42.7% decane) (57.8% cumene) (93.0% butyl
acetate) chloride)
AA-solvent Homogeneous – Homogeneous – Homogeneous Homogeneous –
(32.2% toluene) (23.0% p-xylene) (15.8% decane) (4.3% cumene)
MeOH-solvent Homogeneous Homogeneous – – – – Homogeneous
(26.2% toluene) (63.0% benzene) (67.7% butyl
chloride)
Nhien et al. Biotechnol Biofuels (2017) 10:81 Page 8 of 12

furfural. Figure 8 shows the key design and process Butyl chloride solvent process
parameters of the optimized hybrid extraction/distil- Similar to the toluene process, the cooled feed stream and
lation process using the toluene solvent. Because of the butyl chloride solvent in the butyl chloride process were
high furfural extraction ability, only 44,000 kg/h of tolu- input to an extractor to generate a solvent-rich phase and
ene corresponding to a feed-to-solvent ratio of 1:0.4 was rainate phase. he butyl chloride-rich stream was then
suicient to extract 99.9 wt% of the furfural in the feed. introduced to a distillation column (D2) to isolate the
T he simulation results showed that D1 required furfural at the desired purity of 99 wt% and recover butyl
medium- pressure steam for its reboiler and had an chloride. Figure 9 presents the key design and process
energy con- sumption of 11,354 kW. parameters of the optimized hybrid extraction/distillation

T = 368K
P = 101 kPa
Q = 11,508 kW

43,760 kg/h
50,030 kg/h 1 1 wt% water
Q = 4645 kW 87 wt% toluene Toluene 99 wt% toluene
105,000 kg/h 353K 313K 1 13 wt% furfural 22
2 wt% methanol Rich solvent
90 wt% water Feed D1
2 wt% AA
6 wt% furfural
Toluene E1 Aqueous
26 6270 kg/h
5 98,970 kg/h
44,000 kg/h 1 wt% toluene
T = 433K Furfural 99 wt% furfural
P = 101 kPa
P = 122 kPa
Q = 11,354 kW
Fig. 8 Schematic diagram of hybrid extraction/distillation process using toluene solvent

T = 322K
101 kPa
7802 kW

61,502 kg/h
67,773 kg/h 1
99 wt% butyl chloride
Q = 4645 kW 90 wt% butyl chloride Butyl
105,000 kg/h 353K 9 wt% furfural chloride
313K 1 8
2 wt% methanol
90 wt% water Feed Rich solvent
2 wt% AA D2
6 wt% furfural
Butyl chloride E2 Aqueous
11
62,000 kg/h 6 99,228 kg/h 6270 kg/h
Furfural 99 wt% furfural
P = 101 kPa T = 433K
P = 111 kPa
Q = 8326 kW
Fig. 9 Schematic diagram of hybrid extraction/distillation process using butyl chloride solvent
Nhien et al. Biotechnol Biofuels (2017) 10:81 Page 9 of 12

process using the butyl chloride solvent. As shown in Sustainability analysis


Fig. 4 in solvent screening section, butyl chloride showed Economic and environmental evaluation
a less favorable equilibrium for extracting furfural than In an economic evaluation, the total investment cost
toluene. herefore, a greater amount of butyl chloride (TIC), total annual operating cost (TOC), and TAC of
(62,000 kg/h) was needed to achieve a 99.9 wt% furfural all the processes were calculated based on the Turton
recovery in the extractor. However, the separation in the textbook [26] as described in detail in our previous work
distillation is based on diferences in the mixture volatility. [27]. he Chemical Engineering Index of 556.8 in 2015
T hus, the separation of the butyl chloride–furfural was considered in this work. he column diameters, tray
mix- ture required less energy than the toluene–furfural spacing, and column heights were calculated using the
sepa- ration. The results showed that the butyl chloride tray sizing function in Aspen Plus. he equipment con-
process can save 26.7% of the energy used by the toluene sidered in the investment costs consist of all of the reboil-
process. ers, condensers, coolers, column vessels, and tray stacks.
For the TOC calculation, the costs of a medium-pressure
Benzene solvent process steam at 10 barg and cooling water are 14.19 and 0.354 $/
Figure 10 shows the simplified lowsheet and key GJ, respectively [26].
pro- cess parameters of the optimized hybrid In addition, the total annual CO2 emissions (TCE)
extraction/ distillation process using the benzene were also estimated to assess the environmental impact.
solvent. he aqueous feed and solvent were introduced For the steam reboilers, Gadalla’s modular method was
to an extrac- tor (E3) to produce a benzene-rich applied to calculate the CO2 emissions [28]:
phase at the top and rainate phase at the bottom. Qfuel C%
Next, the benzene- rich stream was input to a
distillation column (D3) to
isolate the benzene and furfural at the desired purities. [CO2]emiss = α, (2)
Among the solvents tested, benzene showed the highest NHV 100
distribution coeficient for furfural extraction. Hence, where NHV is the net heating value of the fuel, and
only 21,000 kg/h of benzene was needed to extract 99.9 C% is the carbon content. For natural gas, the NHV is
wt% of the furfural in the feed, while the necessary 48,900 kJ/kg, and C% is 0.41 kg/kg. he molar mass ratio
amounts of butyl chloride and toluene were 62,000 and of CO2 and C was α = 3.67. In addition, Qfuel denotes the
44,000 kg/h, respectively. The results showed that the amount of fuel used, which was calculated as follows:
benzene process could achieve energy savings of 43.0 Qproc TFTB − T0
and 22.2% compared to the toluene and butyl chloride Qfuel = hproc − 419 , (3)
proc TFTB — Tstack
processes, respectively.

T = 341K
P = 101 kPa
Q = 5576 kW

47,939 kg/h 1 41,669 kg/h


105,000 kg/h Q = 4645 kW 86 wt% benzene Benzene 99 wt% benzene
2 wt% methanol 353K 313K 1 13 wt% furfural 6
90 wt% water Feed Rich solvent
2 wt% AA D3
6 wt% furfural

Benzene E3 Aqueous
9
4 6270 kg/h
42,000 kg/h 99,061 kg/h
T = 433K Furfural 99 wt% furfural
P = 101 kPa 1 wt% benzene
P = 111 kPa
Q = 6476 kW
Fig. 10 Schematic diagram of hybrid extraction/distillation process using benzene solvent
Nhien et al. Biotechnol Biofuels (2017) 10:81 Page 10 of 12

where Qproc is the heat duty requirement of the system, and TCE, respectively, compared to the toluene process.
and λproc (kJ/kg) and hproc (kJ/kg) are the latent heat and Toxicity and eco-toxicity are main undesirable features of
enthalpy of the steam, respectively. he lame, stack, and butyl chloride as a solvent. But if we look at its potential
ambient temperatures were TFTB (2073 K), Tstack (433 K), contribution to a global warming aspect, its signiicant
and T0 (298 K), respectively. CO2 emission reduction efect would give butyl chloride
Table 4 lists the key results for the carbon footprint and most attractive and advantageous feature as a solvent. As
economic performance, including TAC, TIC, and TOC. shown in the present study, the hybrid extraction and dis-
Remarkably, the benzene process showed the best perfor- tillation process using butyl chloride solvent can reduce
mance, achieving savings of 24.1, 60.7, and 57.1% in terms CO2 emission up to 58.3% equivalently compared to the
of TIC, TOC, and TAC, respectively, as well as a reduction traditional distillation process. Its outstanding perfor-
of 43.0% in the carbon footprint compared to the toluene mances on CO2 reduction due to enhanced energy ei-
process. However, because benzene is a human carcino- ciency with the high recovery of 99.71 wt% can be a worth
gen, its use as a solvent in the industry needs to be care- consideration as a possible option of industrial furfural
fully controlled. In recent years, it has been supplanted separation technology. Of course, to take into account its
by other safer solvents. herefore, we here propose butyl potential problem associated with the toxicity, it should
chloride solvent as a good choice for the hybrid purii- be assumed that the process must be designed and oper-
cation process of furfural because its process can save ated to satisfy more strict safety and sealing regulation
17.1, 26.8, 25.9, and 26.7% in terms of TIC, TOC, TAC, like as other processes handling harmful species.

Comparison between hybrid extraction/distillation


Table 4 Comparison of key results for diferent solvent and tradition distillation processes
processes his section discusses how the proposed hybrid extrac-
Properties Solvent tion/distillation process was compared with a traditional
distillation process, which was presented in detail in our
Benzene Toluene Butyl chloride
previous study [4], using the TAC to fully assess its eco-
Condenser duty (kW) 5576 11,508 6995 nomic feasibility. Table 5 lists the key results for the car-
Energy requirement saving in 51.5 – 39.2 bon footprint and costs for the two processes. Note that
condenser (%) in comparison with the traditional distillation process,
Reboiler duty (kW) 6476 11,354 8326 the hybrid process requires an added solvent and metha-
Energy requirement saving in 43.0 – 26.7 nol is not collected as a byproduct. herefore, additional
reboiler (%)
costs for the extraction solvent and the beneit achieved
Investment cost (k$/year) 2761 3640 3018
from selling methanol were considered. To perform a
Investment cost saving (%) 24.1 – 17.1
fair comparison, the same production rate of 50 ktons
Annual operating cost (k$/year) 1969 5004 3663
of furfural per year was assumed for both processes and
Annual operating cost saving (%) 60.7 – 26.8
the relative annual costs (TAC*) were calculated from the
Total annual cost (k$/year) 2380 5547 4113
original TAC. In the distillation process, the annualized
Total annual cost saving (%) 57.1 – 25.9
methanol proit was subtracted from the original TAC,
Total annual CO2 emission (ton/ 13,923 24,411 17,901 while in the hybrid process, the annualized cost of sol-
year)
Total annual CO2 reduction (%) 43.0 – 26.7
vent makeup and the initial solvent cost were added to
TAC and TIC, respectively. he prices of methanol and

Table 5 Comparison of key results between the hybrid and traditional distillation processes and their improved pro-
cesses
Process TCE TIC TOC TAC Methanol Solvent TAC*
(ton/year) (k$/year) (k$/year) (k$/year) proit (k$/year) cost (k$/year) (k$/year)

Distillation process [4] 42,903 4501 8666 9337 −2306 7031


Hybrid process 17,901 3085 3663 4123 1558 5680
Savings (%) 58.3 31.5 57.7 55.8 19.2
Advanced distillation 37,938 4928 7661 8396 −2306 6089
process [4]
Enhanced hybrid process 16,978 2935 3510 3947 1558 5505
Savings (%) 55.2 40.4 54.2 53.0 9.6
Nhien et al. Biotechnol Biofuels (2017) 10:81 Page 11 of 12

butyl chloride taken from global world market data were signiicant efect on the hybrid design. In contrast to the
$0.276/kg [29] and $1.1/kg, respectively [30]. Meanwhile, hybrid process, as the methanol fraction increased, the
the proit from selling methanol was assumed to be 50% TAC* of distillation decreased signiicantly. his means
of the total sales value. that the additional investment and operating costs for
he results showed that the hybrid extraction/distilla- the methanol separation column due to an increase in
tion process using the butyl chloride solvent can produce the methanol lowrate are much lower than the proit
TAC savings of up to 19.2% compared to the traditional from selling methanol. Figure 11 illustrates the beneit of
distillation process. Remarkably, the hybrid process was the hybrid process compared to the distillation process
also more eco-friendly, accounting for a 58.3% reduction with diferent scenarios for the byproduct composition.
in the carbon footprint, which is closely linked to energy Remarkably, the hybrid process is more favorable than the
requirements, compared to the distillation process. distillation process when the methanol fraction of the feed
Furthermore, the results shown in Table 5 also indi- stream decreases and does not produce any beneit. How-
cated that the improved hybrid process, which utilizes ever, the distillation process is a better choice when the
the heat of the distillation bottom stream to preheat its methanol fraction is larger than 3 wt%. It is worth noting
feed, can save up to 9.6% of the TAC as compared to the that the present study used the UNIFAC model to estimate
improved distillation process using integrated and inten- the missing binary parameters of several minor compo-
siied techniques, which was studied by Nhien et al. [4]. nents. he experimental validation is essential to be con-
sidered as a next step for conirming process performance,
Sensitivity analysis leading to more reliable results for real implementation.
In a lignocellulose-based process, breaking down the
complex cellulose–hemicellulose–lignin structure of the Conclusions
lignocellulosic biomass could lead to an inherent uncer- A hybrid extraction/distillation process was proposed to
tainty in the stream composition, which could have a successfully improve the traditional process of furfural
major efect on the suggested design. herefore, a sensi- production, which is energy intensive. Promising solvents
tivity analysis was performed to improve the robustness were suggested through a comprehensive procedure for
of the proposed results. he aqueous feed of the present solvent selection that ranged from solvent screening to
puriication process produced from furfural reactors complete process design. he economic and environmen-
comprises about 6 wt% furfural, 90 wt% water, and 4 wt% tal performances of three processes using the most prom-
byproducts (methanol and AA). In this analysis, the fur- ising solvents (toluene, benzene, and butyl chloride) were
fural and water fractions were ixed, and ive scenarios evaluated for a fair comparison. Overall, benzene and
for the methanol and AA composition of the feed were butyl chloride were found to be the most suitable solvents
explored to determine how the byproduct composition for furfural production because their processes could
afects the proposed design for the furfural puriication achieve TAC savings of 71.3 and 44.7%, respectively, com-
process. Table 6 lists the key results of the sensitivity pared to the toluene process. However, because of ben-
analysis of the hybrid and distillation processes. zene’s obvious toxicity, butyl chloride was proposed as a
In the hybrid process, as the methanol fraction is
increased, the rate of solvent makeup and methanol loss
also increase. As a result, the TAC* gradually increases
from $5.4 million up to $6.2 million with respect to a
methanol fraction increase from 0 to 4 wt%. Note that
changes in the byproduct composition did not have a

Table 6 Sensitivity analysis of methanol and AA composi-


tion in the feed
Methanol–AA fraction (wt%) 0–4 1–3 2–2 3–1 4–0

Hybrid process
Solvent makeup (ton/year) 1240 1320 1416 1568 1880
TAC* (k$/year) 5487 5575 5680 5848 6191
Distillation process
Methanol product rate 0 8356 16,712 25,068 33,424
(ton/year) Fig. 11 Effect of methanol–acetic acid composition on TAC savings
TAC* (k$/year) 8778 7913 7031 6150 5267 of hybrid process
Nhien et al. Biotechnol Biofuels (2017) 10:81 Page 12 of 12

good choice for the hybrid puriication process of furfural. 5. Nhien LC, Long NVD, Lee M. Design and optimization of the levulinic acid
recovery process from lignocellulosic biomass. Chem Eng Res Design.
his suggested hybrid extraction/distillation process can 2016;107:126–36.
save up to 19.2% of the TAC compared to the traditional 6. Malveda MP, Kalin T, Yokose K. Furfural, Report 660.5000 A. CEH Marketing
distillation process. Furthermore, the hybrid process was Research Report; 2011.
7. de Jong W, Marcotullio G. Overview of biorefineries based on co-produc-
also more eco-friendly, accounting for a carbon footprint
tion of furfural, existing concepts and novel developments. Int J Chem
reduction of 58.3% compared to the distillation process. Reactor Eng. 2010;8:1–24.
Interestingly, the proposed sequence was more favorable 8. Vedernikovs N. Process for producing of furfural and acetic acid. Patent
LV11950, Letvia; 1998.
than the distillation process when the methanol fraction
9. Delmas M, Benjelloun-Mlayah B. Organic pulping of cereal straw: from
of the feed stream was <3% and more beneit could be the pilot plant to the first factory. In: 16th European biomass conference
obtained when that fraction decreased. and exhibition—from research to industry and markets; Valencia (Spain).
Firenze: ETA Florence; 2008. pp. 1660–4.
10. de Jong W, Marcotullio G. Process for the production of furfural from
pentoses. US Patent No. US 20120108829 A1; 2012.
Abbreviations
11. Tula AK, Eden MR, Gani R. Process synthesis, design and analysis
CO2: carbon dioxide; CAMD: computer-aided molecular design; NRTL: non-
using a process-group contribution method. Comput Chem Eng.
random two-liquid; HOC: Hayden–O’Connell; VLE: vapor–liquid equilibrium;
2015;81:245–59.
LLE: liquid–liquid equilibrium; LL: liquid–liquid; AA: acetic acid; NHV: net
12. Schilling J, Lampe M, Gross J, Bardow A. 1-stage CoMT-CAMD: an
heating value; TIC: total investment cost; TOC: total annual operating cost; TCE:
approach for integrated design of ORC process and working fluid using
total annual CO2 emissions; TAC: total annual cost; TAC*: final total annual cost;
PC-SAFT. Chem Eng Sci. 2017;159:217–30.
UNIFAC: Universal Quasichemical Functional-Group Activity Coefficients.
13. Struebing H, Obermeier S, Siougkrou E, Adjiman CS, Galindo A. A QM-
CAMD approach to solvent design for optimal reaction rates. Chem Eng
Authors’ contributions
Sci. 2017;159:69–83.
LCN and NVDL contributed equally to this work, designed the study, carried
14. Gebreslassie BH, Diwekar UM. Efficient ant colony optimization for
out the process simulations, and wrote the paper. SK advised process design.
computer aided molecular design: case study solvent selection problem.
ML conceived the core concepts for the research and advised academically. All
Comput Chem Eng. 2015;78:1–9.
authors read and approved the final manuscript.
15. Hostrup M, Harper PM, Gani R. Design of environmentally benign pro-
cesses: integration of solvent design and separation process synthesis.
Author details
1 Comput Chem Eng. 1999;23(10):1395–414.
School of Chemical Engineering, Yeungnam University, Gyeongsan 712-749,
16. Nhien LC, Long NVD, Kim S, Lee M. Design and assessment of hybrid
South Korea. 2 Green Material and Process Group, Korea Institute of Industrial
purification processes through a systematic solvent screening for the
Technology, Cheonan 31056, South Korea.
production of levulinic acid from lignocellulosic biomass. Ind Eng Chem
Res. 2016;55(18):5180–9.
Acknowledgements
17. Karunanithi AT, Achenie LEK, Gani R. A new decomposition-based
Not applicable.
computer-aided molecular/mixture design methodology for the
design of optimal solvents and solvent mixtures. Ind Eng Chem Res.
Competing interests
2005;44(13):4785–97.
The authors declare that they have no competing interests.
18. Harper PM, Gani R. A multi-step and multi-level approach for computer
aided molecular design. Comput Chem Eng. 2000;24(2):677–83.
Availability of data and materials
19. Aspentech. Aspen physical property system: physical property methods; 2013.
All data generated or analyzed during this study are included in this published
pp. 27–8.
article.
20. Cheremisinoff NP. Industrial solvents handbook, revised and expanded.
Boca raton: CRC Press; 2003.
Consent for publication
21. Smallwood IM. Solvent recovery handbook. 2nd ed. Great Britain: Black-
All authors approved the manuscript.
well Science; 2002.
22. Hanson TC. Handbook of solvent extraction. New York: Wiley; 1983.
Funding
23. Demesa AG, Laari A, Tirronen E, Turunen I. Comparison of solvents for the
This work was supported by a National Research Council of Science & Technol-
recovery of low-molecular carboxylic acids and furfural from aqueous
ogy (NST) grant by the Korean government (MSIP) (No. CAP–11–04–KIST). This
solutions. Chem Eng Res Design. 2015;93:531–40.
study was also supported by Priority Research Centers Program through the
24. Ruiz F, Gomis V, Botella RF. Extraction of ethanol from aqueous solution.
National Research Foundation of Korea (NRF) funded by the Ministry of Educa-
1. Solvent less volatile than ethanol: 2-ethylhexanol. Ind Eng Chem Res.
tion (2014R1A6A1031189).
1987;26(4):696–9.
25. Offeman RD, Stephenson SK, Robertson GH, Orts WJ. Solvent extrac-
Received: 27 October 2016 Accepted: 22 March 2017
tion of ethanol from aqueous solutions. I. Screening methodology for
solvents. Ind Eng Chem Res. 2005;44(17):6789–96.
26. Turton R, Bailie RC, Whiting WB, Shaeiwitz JA, Bhattacharyya D. Analysis,
synthesis, and design of chemical processes. 4th ed. Upper Saddle River:
Prentice Hall; 2012.
References 27. Long NVD, Minh LQ, Nhien LC, Lee M. A novel self-heat recupera-
1. Werpy T, Petersen G. Top value added chemicals from biomass. Volume tive dividing wall column to maximize energy efficiency and column
I—results of screening for potential candidates from sugars and synthesis throughput in retrofitting and debottlenecking of a side stream column.
gas. Golden: National Renewable Energy Laboratory (NREL); 2004. Appl Energy. 2015;159:28–38.
2. Hoydonckx HE, Rhijn WMV, Rhijn WV, de Vos DE, Jacobs PA. Ullmann’s 28. Gadalla MA, Olujic Z, Jansens PJ, Jobson M, Smith R. Reducing CO2 emis-
encyclopedia of industrial chemistry, vol. 16. 7th ed. 2012. sions and energy consumption of heat-integrated distillation systems.
3. Zeitsch KJ. The chemistry and technology of furfural and its many by- Environ Sci Technol. 2005;39(17):6860–70.
products. Netherlands: Elsevier Science; 2000. 29. https://www.methanex.com/our-business/pricing. Accessed 20 Oct 2016.
4. Nhien LC, Long NVD, Kim S, Lee M. Design and optimization of intensified 30. https://www.zauba.com/import-n+butyl+chloride-hs-code.html.
biorefinery process for furfural production through a systematic proce- Accessed 20 Oct 2016.
dure. Biochem Eng J. 2016;116:166–75.

You might also like